When parents hear that artificial intelligence is entering the curriculum in Class 3, the reaction is usually some mixture of curiosity and unease. Eight-year-olds and machine learning do not obviously belong in the same sentence.
The reassuring part is what the curriculum actually says. For Classes 3 to 5, most of it happens without a screen at all — puzzles, patterns, games, and step-by-step reasoning folded into subjects your child already studies. The letters "AI" do not appear until Class 6. And when they do, a full quarter of that teaching is about ethics, privacy and bias rather than technology.
Here is what CBSE has introduced, taken from the framework document itself.
What Was Notified
Through Circular Acad-15/2026 dated 1 April 2026, CBSE launched a Computational Thinking and Artificial Intelligence curriculum framework for Classes III to VIII, beginning with the 2026-27 session. A week later, Circular Acad-18/2026 released students' and teachers' handbooks for all six classes, free to download.
The handbooks are explicitly "not designed as textbooks, but rather as a companion to the NCERT Mathematics textbooks." That single line tells you the shape of the whole thing: this is not a new subject bolted onto an already full timetable. It is a way of thinking, taught inside the subjects that are already there.
Classes 3 to 5: Mostly Without a Screen
The preparatory stage gets 50 hours a year, built around four skills:
- Abstract thinking — visualising three-dimensional objects, rotations, symmetry
- Pattern recognition — spotting what changes and what repeats
- Decomposition — breaking a problem into smaller pieces
- Algorithmic thinking — following and building step-by-step rules
Difficulty climbs gently across the three years. A Class 3 child works with simple patterns of one or two changes; by Class 5 the clues are interconnected and have to be held together at once.
Crucially, most of this is unplugged — hands-on activities, games and puzzles rather than time in front of a computer. The framework does introduce basic block-based coding such as Scratch, but the centre of gravity at this stage is a child reasoning aloud with a worksheet, not a child on a device. The work is woven into Mathematics and The World Around Us, which is why the handbooks track the NCERT chapters.
Classes 6 to 8: Where AI Actually Begins
The middle stage gets 100 hours a year, divided three ways:
- 40 hours of advanced computational thinking
- 20 hours of introductory AI concepts
- 40 hours of interdisciplinary projects
The AI component is organised into four units: Introduction to AI and Everyday Examples; Basic Data Concepts; Simple Pattern Recognition and Decision Making; and Ethics and Digital Responsibility.
The progression across the three years is specific:
Class 6 covers the basic ideas of AI and where it already appears in daily life, including the difference between supervised and unsupervised learning.
Class 7 introduces prediction — regression, which estimates a number from patterns in past data, and classification, which sorts things into groups — along with AI domains such as computer vision and natural language processing.
Class 8 works through the AI project cycle: define the problem, collect data, test AI tools, reflect and improve. Students research where bias enters a dataset and apply basic strategies for fairness.
The framework recommends free and open-source tools throughout, "to ensure accessibility and adaptability for all institutions" — a deliberate choice that keeps schools from being locked into any one company's platform.
A Quarter of the AI Teaching Is About Ethics
One of the four AI units is Ethics and Digital Responsibility, and the theme runs well beyond it.
Class 6 students learn about digital footprints, privacy, and basic internet safety — secure passwords, safe online behaviour. Class 7 students learn to "explain what bias in AI means, and identify situations where AI can give unfair results." Class 8 extends that to accountability, privacy, and the social impact of how AI gets deployed.
The line in the framework that stayed with us is this one: regular instruction in AI ethics teaches students to verify information, detect bias, and recognise "when AI reliance crosses into cognitive delegation or cognitive over-reliance."
That is a genuinely thoughtful piece of curriculum design, and it is the part we would most want parents to notice.
What We Think
We will be direct about where we think the risk lies, because it is not the one most people worry about.
The danger is not that children will learn too much about AI too young. The framework is careful, age-appropriate, and mostly screen-free for the younger classes. The danger is that a curriculum like this gets taught as a computer class — a room, a timetable slot, a set of software to demonstrate — when its entire purpose is to change how a child approaches an unfamiliar problem in any subject.
A child who can break a problem into parts, notice a pattern, and describe a method step by step is better at mathematics, better at science, and better at reading a difficult passage. The computational thinking is the point. The AI is what makes it urgent.
The second thing we would say to parents: the ethics component is not a formality to be rushed through in the last week of term. A generation that will have answers available on demand for the whole of their working lives needs to be taught, deliberately, when to stop and check. We intend to teach it that way.
Our Atal Tinkering Lab already gives students a place to build and test ideas rather than only read about them, and this curriculum has an obvious home there.
How It Is Assessed
There is no board examination attached to this. The framework asks schools to move away from "traditional summative assessment to continuous, formative, and competency-based assessment."
In practice that means written tests with puzzles and reasoning questions, group activities, teacher observation journals, thematic projects, and discussions about ethical dilemmas. Classes 6 to 8 add practical work. What is being judged is the thinking process, not whether a final answer matched a key — the handbooks tell teachers explicitly to prioritise "the thinking process over correctness."
Parents used to marks out of a hundred may find this unfamiliar. It is a deliberate choice, and we think it is the right one for a subject whose whole value lies in how a child reasons.
What Is Not Yet Settled
Two honest caveats.
The circular does not state whether the framework is mandatory or optional for affiliated schools, or prescribe how the hours must sit in the timetable. Schools are working that out in the first year, and so are we.
Second, this framework covers Classes III to VIII only. Reports of computational thinking and AI arriving in Classes 9 to 12 are not, as of now, backed by any CBSE circular. We will write about that when a circular exists and not before.
If you would like to know how these hours will look in your child's week, please contact the school office and we will walk you through the timetable as it settles.
Sources. This post is based on CBSE's own documents:
- Circular Acad-15/2026, 1 April 2026 — launch of the CT and AI curriculum framework for Classes III–VIII
- CBSE Curriculum Framework: Computational Thinking and Artificial Intelligence, Classes III–VIII, 2026-27 — hours, units, class-wise progression, assessment and ethics
- Circular Acad-18/2026, 9 April 2026 — students' and teachers' resource books
Where a detail was not stated in an official CBSE document, we have said so rather than reproducing it from secondary reporting.